2010 Oxford Business & Economics Conference Program ISBN : 978-0-9742114-1-9

REDUCTION OF BLASTING RELATED MINING PROCESS VARIABILITY

Dave Lilly, P.E., MBA Baron Fidler

Senior Engineering Consultant General Manager/Technical Resources
Dyno Consult, a Division of Dyno Nobel Inc. Dyno Consult, a Division of Dyno Nobel Inc.
Pawleys Island, South Carolina Salt Lake City, Utah

Larry Mirabelli Ran Tamir

Senior Project Manager Mine Superintendent
Dyno Consult, a Division of Dyno Nobel Inc. Lafarge North America
Roswell, Georgia Ravena, New York

John Cory Bill Hissem

Blasting Specialist North America Senior Mining/Application Engineer
Lafarge North America Sandvik Mining and Construction
Ottawa, Ontario Appleton, Wisconsin

INTRODUCTION

In the past few years advances in blasting technology have significantly reduced
the variability in blasting operations. There are many practical methods to reduce
variability in blasting operations including highly accurate electronic detonators to
control the timing of blasts, more precise drills to control the geometric layout of
blast, and variable property bulk loaded explosives to control blast properties. In
addition, there is a growing knowledge in the industry that blasts must be
properly laid out and prior blasts cleaned up. The use of accurate GPS
equipment to control blast geometry is a quantum leap forward. In addition drill
borehole accuracy now can be accurately monitored as well as face angles.

Fragmentation analysis of final blast results and elaborate blast databases allow
the adjustment of future blasts to produce accurate size and productivity results.

This paper presents a method of determining the variability of mining results from
blasting and a method to determine the major causes of blasting related
variability in the mining process. Mining results could be blasted product size,
productivity, cost, vibration or air blast or a combination of all. The authors will
introduce concepts to reduce variability.

- Principal component analysis or regression analysis to determine

important factors in variability

- The regression of important factors that determine the standard

deviation and mean of the critical variables

- Inclusion of the “Z Table” to estimate percentage in specification, and

- The use of the Excel solver program to formulate a linear programming

constrained analysis of the variability to find the best solution.

June 28-29, 2010 1

St. Hugh’s College, Oxford University, Oxford, UK
2010 Oxford Business & Economics Conference Program ISBN : 978-0-9742114-1-9

An example will be illustrated using a simple one variable (powder factor) model
to maximize the percentage of blasted rock within established dimensions. The
model is developed from fragmentation distribution information related to powder
factor from the Lafarge Ravena, NY Quarry.

THE PROBLEM

Blasting is normally a separate process from other mining processes such as

crushing, milling, and final mineral processing. An outside contractor normally
supplies explosives and integration of blasting into the mining process is difficult
at best. The relationships are further distorted by variations in geologic
conditions.

Never the less, the improvements in blasting variability and results as detailed in
the introduction have made the blasting process much more reproducible. This
paper introduces an algorithm to better control and maximize the amount of
aggregate within a given size range.

THE DATA

Blast reports and blasting databases can be used to determine the factors
affecting blast results. A typical blast report and portions of a blasting database
are shown below.

Quarry Blast Date QBench Blast # Tons Hole Diameter BXS holes
Ravena: CM 33-CM-10-
Ravena Aug 25, 2009 Main Pit 2009 72505.26 6.5 18X20 34
Ravena: CM 31-CM-09-
Ravena Aug 04, 2009 Main Pit 2009 65869.32 6.5 16X19 37
Ravena: CM 29-CM-08-
Ravena Jul 21, 2009 Main Pit 2000 103098.9 6.5 16X19 58
Ravena: CM 26-CM-07-
Ravena Jun 25, 2009 Main Pit 2009 68026.3 6.5 17X19 34
Ravena: CM 19-CM-06-
Ravena May 20, 2009 Main Pit 2009 88834.09 6.5 17X19 45
Ravena : CM 15CM 06-
Ravena Apr 09, 2009 West 2009 97318 6.5 17X19 48
Ravena: CM
Ravena Apr 02, 2009 Main Pit 13-09 99345.98 6.5 17X19 49
Ravena: CM
Ravena Mar 25, 2009 Main Pit 11-09 70001.28 6.5 17X19 41
Ravena: CM
Ravena Feb 26, 2009 Main Pit 07-09 62427.36 6.5 17X19 38
Ravena: CM
Ravena Feb 11, 2009 Main Pit 04-09 64265.34 6.5 17X19 33

June 28-29, 2010 2

St. Hugh’s College, Oxford University, Oxford, UK
 2010 Oxford Business & Economics Conference Program ISBN : 978-0-9742114-1-9

In addition, video analysis of the broken rock mass can quantify the degree of
fragmentation and specifically the mean size and standard deviation of the
blasting size results from a blast.

THE SOLUTION

Once enough data is accumulated from blasts and size distribution analysis, the
relationship between blasting parameters and the mean and standard deviation
of the size distribution can be determined. In this simple example, only powder
factor (tons of rock per lb of explosive) is used as a determining variable. Other
variables such as pattern area can be included in the model also. Principal
component analysis or multivariable curvilinear regression can be used to
determine the significant variables.

The relationships between powder factor and mean size and size standard
deviation are shown below.

AGGREGATE AGGREGATE

9
SIZE STANDARD DEVIATION

10
y = 4.8795x - 2.5093
MEAN BLOCK SIZE IN

9 8
R2 = 0.4073
8 7

7 6
y = 8.7486x - 10.714
6 5
R2 = 0.6213
5 4

4 3
1.6 1.7 1.8 1.9 2 2.1 2.2 1.6 1.7 1.8 1.9 2 2.1 2.2
POWDER FACTOR - TONS/LB POWDER FACTOR - TONS/LB
The relationships from the charts can be expressed as equations:

June 28-29, 2010 3

St. Hugh’s College, Oxford University, Oxford, UK
2010 Oxford Business & Economics Conference Program ISBN : 978-0-9742114-1-9

Mean Size = -2.5093 + 4.8795 x Powder Factor

Standard Deviation = -10.714 +8.7486 x Powder Factor

In addition, the data from a standard Z distribution (The “Z value” is the distance
between a value and the mean in terms of standard deviations) can be put in the
form of an equation and used to determine the percentage of rock lying within the
upper and lower bounds determined by the standard deviation and mean. The Z
distribution and resulting equation is shown below:

0.6

0.5

0.4
AREA

0.3 y = 0.0162x 3 - 0.1556x 2 + 0.4905x - 0.0109

R2 = 0.9995
0.2

0.1

0
0 1 2 3 4 5
Z

The above equations can be put into an interactive spreadsheet as shown below.
The Excel Solver from the tools menu is activated and the amount of rock
between the upper and lower limit is maximized based upon changes in powder
factor and subsequent changes in mean and standard deviation.
Powder
Factor
Tons/Lb
Solution 1.6 Calculated Calculated
Minimum Maximum
Block Size Mean Std. Dev. Size Size
5.2979 3.28376 -4.55338 15.14918

Range Limits Upper Lower Z1 Z2

10 2 1.431926 1.004306
10 2
Maximize % in Range 76.1%

0.14

0.12

0.1

0.08 Series1
UPPER LIMIT
0.06 LOWER LIMIT
0.04

0.02

0
The results were non
-10 intuitive
-5 but
0 upon
5 reflection
10 15since
20 both mean and standard

deviation increased with increasing powder factor it makes sense. The graph

June 28-29, 2010 4

St. Hugh’s College, Oxford University, Oxford, UK
2010 Oxford Business & Economics Conference Program ISBN : 978-0-9742114-1-9

below shows the change in distribution as the powder factor is increased or

lowered. The optimum powder factor is 1.6 tons/lb and has 76.1% of the
produced aggregate within the upper and lower limits. For reference, the 2.1
tons/lb has a 39.1% of aggregate within the acceptable range.

0.14

0.12

0.1
1.6 tons/lb
0.08 UPPER LIMIT

0.06 LOWER LIMIT

2.1 ton/lb
0.04

0.02

0
-10 0 10 20 30
SIZE - INCHES

THE FUTURE

The simple model above can be expanded with other blasting variables. In
addition, the techniques can be used to maximize productivity, minimize cost and
to understand the variability in air blast and vibration.

ACKNOWLEDGEMENTS

The authors would like to thank the following for their assistance and help with
this project: John Truszkowski, Dave Beckwith, Dave Warpula, Dyno Nobel; and
Chris Luchkiw, Lafarge.

REFERENCES

Bremer, D., Ethier, R., and Lilly, D. (2007), Factors Driving Continuous Blasting
Improvement at the Lafarge Ravena Plant. International Society of Explosives
Engineers –Annual Conference on Blasting Technique.

Cory, J., Hissem, B., and Lilly, D. (2009), The Use of Principal Component
Analysis to Integrate Blasting Into the Mining Process. Oxford Business and
Economic Conference.

Cory, J., Dann, M., and Lilly, D. (2008), The Enterprise Solution to Integrating
Blasting Into the Mining Process, Oxford Business and Economic Conference.

June 28-29, 2010 5

St. Hugh’s College, Oxford University, Oxford, UK
2010 Oxford Business & Economics Conference Program ISBN : 978-0-9742114-1-9

Deshpande, P., Tantalean, Unifying Framework for Six Sigma and Process
Control. Hydrocarbon Processing, June 2009.

Dann, M., Smith, C., and Lilly, D. (2007), Aggregate Size Optimisation Program
at the Lafarge Marblehead Plant. 7th Large Open Pit Mining Conference, Perth,
Australia.

Lilly, David P. (2007), A Statistical Approach to Integrating Blasting Into the Mining
Process, Oxford Business and Economic Conference.

Mirabelli, L. (2000). Mining Automation Program

Mirabelli, L., Lilly,D. (2008) A Blaster’s Tool to Measure Fragmentation.

International Society of Explosives Engineers –Annual Conference on Blasting
Technique.

Mirabelli,L.; Hissem, B.; and Veltrop, G. (2009). Missouri Quarry Productivity

Improvement – Casework. International Society of Explosives Engineers –Annual
Conference on Blasting Technique.

Tamir, R. (2007).Tools to help Improve Operational Performance and Control

Operating Costs-or Reducing the Quesstimation Factor

June 28-29, 2010 6

St. Hugh’s College, Oxford University, Oxford, UK